Fuel Indexes: A Novel Method for the Evaluation of Relevant

Nov 30, 2011 - *Telephone: +43-(0)-316-481300-73. ..... appliances: Effect of blending on slag formation in the bottom ash and pre-evaluation options...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/EF

Fuel Indexes: A Novel Method for the Evaluation of Relevant Combustion Properties of New Biomass Fuels Peter Sommersacher,*,† Thomas Brunner,†,‡,§ and Ingwald Obernberger†,‡,§ †

BIOENERGY 2020+ GmbH, Inffeldgasse 21b, A-8010 Graz, Austria Institute for Process and Particle Engineering, Graz University of Technology, Inffeldgasse 21a, A-8010 Graz, Austria § BIOS BIOENERGIESYSTEME GmbH, Inffeldgasse 21b, A-8010 Graz, Austria ‡

ABSTRACT: The increasing demand for biomass fuels leads to the introduction of new biomass fuels into the market. These new biomass fuels (e.g., wastes and residues from agriculture and the food industry, short rotation coppices, and energy crops) are usually not well-defined regarding their combustion behavior. Therefore, fuel characterization methods with a special focus on combustion-related problems (gaseous NOx, HCl, and SOx emissions, ash-melting behavior, and PM emissions) have to be developed. For this purpose, fuel indexes are an interesting option. Fuel indexes are derived from chemical fuel analyses and are checked and evaluated regarding their applicability by measurements performed at lab- and real-scale combustion plants for a large variety of fuels. They provide the possibilities for a pre-evaluation of combustion-relevant problems that may arise from the use of a new biomass fuel. A possible relation to describe the corrosion risk is, for instance, the molar 2S/Cl ratio. The N content in the fuel is an indicator for NOx emissions, and the sum of the concentrations of K, Na, Zn, and Pb in the fuel can give a prediction of the aerosol emissions, whereas the molar (K + Na)/[x(2S + Cl)] ratio provides a first indication regarding the potential for gaseous HCl and SOx emissions. The molar Si/K ratio can supply information about the K release from the fuel to the gas phase. The molar Si/(Ca + Mg) ratio can give indications regarding the ash-melting temperatures for P-poor fuels. For Prich fuels, the (Si + P + K)/(Ca + Mg) ratio can be used for the same purpose. The fuel indexes mentioned can provide a first pre-evaluation of combustion-relevant properties of biomass fuels. Therefore, time-consuming and expensive combustion tests can partly be saved. The indexes mentioned are especially developed for grate combustion plants, because interactions of the bed material possible in fluidized-bed combustion systems are not considered.

1. INTRODUCTION, BACKGROUND, AND SCOPE The general increase in energy demand, higher costs, and oncoming depletion of fossil fuels are reasons for considerable research on renewable energies and raw material resources. Biomass energy can be an important factor in reducing greenhouse gas emissions and energy source independence.1 To increase the biomass use in the energy sector, products such as wastes and residues from agriculture and the food industry, as well as short-rotation plants, such as poplar, and energy crops, such as Miscanthus, Arundo donax, and grasses, can be used. These new biomass fuels gain rising interest for use in biomass combustion systems. As a result, new biomass fuels are of relevance for furnace and boiler manufacturers as well as utilities. According to the increasing prices for conventional biomass fuels (e.g., wood), new options for heat and power generation from cheap feedstocks have to be identified to increase the economic efficiency of biomass heating and combined heat and power (CHP) plants. The introduction of new biomass fuels may also create new employment opportunities in rural areas and, therefore, also contribute to the social aspect of sustainability.2,3 For the past decade, agricultural studies throughout Europe have been focusing on introducing new nonfeed crops with a perspective for industrial applications. Perennial rhizomatous grasses, such as switchgrass (Panicum virgatum), Miscanthus spp., giant reed (A. donax), and red canarygrass (Phalaris arundinacea), have been considered as the most promising energy crops for Europe. The employment of such raw © 2011 American Chemical Society

materials contributes to reduce anthropogenic CO2 emissions. Also, the situation of the agricultural sector in the European Union (EU) can be improved. This sector is characterized by food surplus, which has led to a policy of setting land aside. Therefore, introducing alternative nonfood crops as energy crops can represent a new opportunity for the population of rural areas. Perennial grasses, such as Miscanthus and giant reed, show some ecological advantages. They require limited soil management and a low demand for nutrient inputs. Also, further restoration of degraded land may be possible.4 Short-rotation coppices (SRCs) are fast-growing tree species, which are used to produce biomass as a renewable energy source. The most common SRCs are willow and poplar, which have been shown to be viable alternatives to fossil fuels.2 Agricultural residues, such as olive stones, are already used as fuels for combustion processes; however, a great further potential of these fuels exist, which is presently not used but disposed.5 However, because new biomass fuels are not well-defined yet regarding their combustion behavior, fuel characterization with a special focus on combustion-related fuel properties is a first important step for their introduction. Possible ash-related problems (ash melting on grates, bed agglomeration in fluidized-bed combustion systems, deposit formation and Received: August 29, 2011 Revised: November 25, 2011 Published: November 30, 2011 380

dx.doi.org/10.1021/ef201282y | Energy Fuels 2012, 26, 380−390

Energy & Fuels

Article

sintering or slagging, fouling, and high-temperature corrosion on biomass fuels. Far less empirical correlations or fuel indexes predicting combustion-related problems have thus far been found in the literature for biomass. For fluidized-bed systems, the indexes (K + Na)/(2S + Cl), (K + Na + Si)/(Ca + P + Mg), and K/Si have been investigated.7 The index (Na + K)/(2S + Cl) is based on the general observation that the initial gas-phase alkali concentration is attributable to the alkali concentration in the fuel. A molar (Na + K)/(2S + Cl) ratio in the fuel >1 reflects the chance that the excess alkali over the sum of S and Cl would stay in the bed and may react with silicates. The formation of alkali silicates is often observed and leads to sintering or defluidization. The index (K + Na + Si)/(Ca + P + Mg) is applied to describe the probability of the formation of a nonsticky outer coating on a bed material particle. For a ratio smaller than 1, the refractory outer coating is likely to form and prevent the sintering of coatings. Also, a third agglomeration indicator, the molar K/Si ratio, was introduced, where values >1 indicate that the occurrence of agglomeration is more likely a result of melt-induced sintering than the sintering of coating. On the basis of results of test runs in a residential biomass pellet boiler with 12 different fuels, which were combusted, the slagging tendencies and the K retention have been investigated. The molar amounts of Si-(Cl + Ca + Mg) in the fuel should describe the K retention in residual ash (slag and bottom ash). The second correlation, the molar amounts of Si-(Cl + Ca + Mg) in the residual ash, should describe the K retention in the slag and the slagging tendency. There is a trend reported for the second correlation, where with increasing value of this index, the K retention in the slag and the slagging tendency increase.8 It is assumed that the elements Cl, Ca, and Mg may increase the K release, whereby higher Si contents in the fuel favor the K retention in the ash and slag. The molar Si/(Ca + Mg) ratio has been reported to predict the slagging tendency for residential heating appliances, where with rising value, slagging increases.9 Although the indexes of (K + Na)/(2S + Cl), (K + Na + Si)/ (Ca + P + Mg), and K/Si were developed for fluidized-bed systems, where interactions of the bed material influence the combustion behavior, some fundamental considerations may also be used within this work focusing on fixed-bed systems. The aim of this work was the evaluation of the applicability of already defined fuel indexes as well as the development of new fuel indexes, which can be applied to predict combustionrelated problems. This work is based on data from lab, pilot, and field tests performed with a broad spectrum of biomass fuels at fixed-bed combustion systems.

corrosion, and particulate emissions) as well as problems regarding NOx, SOx, and HCl emissions are thereby the main focus points. With state-of-the-art fuel characterization methods, these issues cannot be sufficiently covered, and therefore, new and advanced fuel characterization methods are needed. One option to gain quick first indications regarding the problems mentioned is the development, evaluation, and application of fuel indexes as one advanced biomass fuel characterization tool. 1.1. Background and Scope. The fireside behavior of minerals in coal was first investigated more than 100 years ago. After the beginning of burning high-sulfur coals rich in pyrites (FeS2) on grates, it was quickly recognized that pyrite was responsible for the formation of clinkers. For the first time, a combustion-related problem was directly related to a specific mineral species. It was understood that iron occurring in highsulfur coals acted as a fluxing agent, lowering the melting temperature of quartz and clays found in coal. These early problems associated with slagging were the trigger for the development of empirical correlations for the prediction of slagging tendencies. The correlations describe the relationship between the melting temperature of slag and the proportional distribution between basic and acidic constituents.6 Afterward, the increasing electricity demand since the 1950s as well as the ambition to increase the electric efficiency resulted in continuously increasing steam temperatures. Also, the use of more problematic coals resulted in further fireside problems, such as ash sintering or slagging, fouling of convective heat recovery surfaces, and high-temperature corrosion. Therefore, various empirical correlations especially to predict the ash-sintering behavior of coal have been developed.6 In a pre-evaluation step of this work, existing empirical correlations for coals were tested regarding their applicability for biomass fuels. It was seen that indexes developed for coal cannot be applied for biomass. Because coal is a type of sedimentary rock, minerals can occur disposed as tiny inclusions within macerals, layers, nodules, fissures, and rock fragments. Thick layers and abundant nodules are removed by standard preparation facilities. The thinner layers and nodules stay in the coal and typically consist of aluminum silicates (clay, illite, kaolinite, feldspar) and silicon oxide (quartz). The alkali metals (typically Na) are bound in this aluminum silicate structure and occur in minor concentrations. In biomass fuels, in contrast, K exists in high concentrations, and its chemical binding in the fuel matrix is different from coal. K occurs as highly mobile ion in biomass fuels. Si as a further example is introduced in the plants by absorption of silicate acid from the soil. Si is deposited as a hydrated oxide usually in an amorphous form but occasionally in crystalline form. Dependent upon the biomass species, the major elements responsible for the ash chemistry can roughly be categorized in low-Si-/high-Cacontaining fuels, e.g., wood and woody biomass and high-Si-/ low-Ca-containing fuels, e.g., herbaceous biomass. It can generally be said that coals contain higher amounts of S mostly in the form of FeS2. In biomass plants, S exists as sulfates or organic sulfur. The amount of Cl in coal depends upon the coal type and exists predominantly as inorganic alkali chlorides and a smaller amount of unspecific organic chlorides. In biomass, Cl appears as a chloride ion, where its concentration is closely related to the nutrient composition of the soil.6 The difference in occurrence and in chemical binding of certain elements explains the non-applicability of empirical coal correlations for

2. METHODOLOGICAL APPROACH 2.1. Fuels Investigated. Biomass fuels from five different biomass categories, which are typical for different climate zones, were used within this study: (1) wood and woody biomass (WWB): (i) beech woodchips from Austria, (ii) spruce wood chips from Austria, (iii) spruce pellets from Austria, (iv) bark of softwood, (v) waste wood, and (vi) torrefied softwood; (2) SRC: (i) poplar from Austria; (3) herbaceous and agricultural biomass (HAB): (i) Miscanthus pellets from Germany, (ii) wheat straw from Austria, (iii) olive kernels from Greece, (iv) maize spindel pellets (maize residues) from Austria, and (v) grass pellets; (4) others: (i) sewage sludge from Austria; and (5) industrial biomass residues: (i) decanter and rapeseed press cake and (ii) residues of starch production. The group WWB represents the conventional biomass fuels, whereas the remaining groups represent new biomass fuels. 381

dx.doi.org/10.1021/ef201282y | Energy Fuels 2012, 26, 380−390

Energy & Fuels

Article

Figure 1. Scheme of the lab-scale reactor, including measurement setup (left), definition of common test conditions for all experiments (middle), and projection of the lab-scale reactor fuel bed to the fuel layer in a grate furnace (right).10,11 2.2. Methodology. Proximate and ultimate analyses of the fuels investigated (ash content, contents of N, S, Cl, and major and minor ash-forming elements) are applied as the basis for the work presented. The analytical methods are described in section 3. From the results of these analyses, fuel indexes are calculated and evaluated. They are defined by considering the physical behavior and chemical reactions of dedicated elements during biomass combustion, known interactions of different ash-forming elements during thermal biomass conversion, and correlations and experiences gained from pilot- and real-scale combustion as well as lab-scale combustion tests with conventional and new biomass fuels. Data derived from combustion tests performed at BIOENERGY 2020+, the Institute for Process and Particle Engineering, Graz University of Technology, and BIOS BIOENERGIESYSTEME GmbH, Graz, Austria, were considered. 2.3. Pilot- and Real-Scale Combustion Tests. All combustion plants have geometrically separated primary and secondary combustion zones and, thus, enable an efficient air staging. The primary air ratio (amount of primary air/stoichiometric amount of air) is typically below 1 (0.6−0.9), and the overall air ratio applied is between 1.4 and 1.6. The plants are also equipped with flue gas recirculation, which ensures that the bed temperature is kept in a range of 900−1100 °C. The real-scale combustion plants investigated are also grate-fired combustion systems with nominal thermal boiler capacities between 0.5 and 110 MW. Also, these boilers are equipped with air-staging technology, and most of them are equipped with flue gas recirculation. Grate combustion plants representing the present state-of-the-art were chosen, and only comparable combustion setups were considered within the evaluation. During the pilot- and real-scale combustion tests, the following data were collected: (1) Fuel sampling and subsequent analyses: (i) moisture content, ash content, and chemical composition. (2) Aerosol and fly ash sampling: (i) For aerosol emissions, low-pressure-cascade impactors (Berner-type low-pressure impactors) were used. (ii) For total fly ash emissions, the gravimetric method according to VDI 2066 was used. (3) Deposit sampling with deposit probes. (4) Emission measurements concerning CO, CO2, NO, NOx, O2, SOx, and HCl: (i) CO, CO2, and O2 were measured with a multi-component gas analyzer (Rosemount NGA 2000). This device uses non-dispersive infrared measurement technique (NDIR) for CO and CO2 analyses and paramagnetism as a measurement principle for O2 detection. (ii) A nitric oxide analyzer (ECO Physis CLD 700 El ht), which uses the principle of chemiluminescence, has been used for NO and NOx detection. (iii) The SOx and HCl emissions were performed by discontinuous measurements according to VDI 3480, Sheet 1. (5) Determination of the furnace temperatures. (6) Sampling of all relevant ash fractions (bottom ash, furnace ash, cyclone ash, filter fly ash, and aerosols) and subsequent analyses: (i) chemical composition. (7) Recording of all relevant operating parameters (O2 content in the flue gas, furnace temperature, load, combustion air supply, etc.)

The methods used for fuel and ash analyses are described in section 3. On the basis of the analysis results and measurement data, energy, mass, and element balances over the plant were calculated. Moreover, recovery rates for all ash-forming elements considered were evaluated to ensure the quality of the data. Only test runs with recovery rates >90% were used for further evaluation. 2.4. Lab-Scale Reactor Tests. In addition, results from test runs with a lab-scale reactor especially designed for the investigation of the thermal decomposition behavior of biomass fuels have been considered. This lab-scale reactor consists of a cylindrical retort (height, 35 cm; inner diameter, 12 cm), which is heated electrically and controlled by two separated proportional−integral−derivative (PID) controllers (see Figure 1, left). The fuel is put in a cylindrical holder of 100 mm height and 95 mm inner diameter. The material of the reactor wall and sample holder is silicon carbide, which is inert under reducing and oxidizing conditions; therefore, the wall does not react with the fuel, ash, and flue gas. The mounting and vessel for the fuel bed are placed on the plate of a scale. The scale is mechanically separated from the retort by a liquid sealing (synthetic thermal oil: Therminol 66). The scale is used to determine the weight loss of the sample. With this setup, it is possible to continuously measure the mass reduction of the sample during drying, pyrolysis, gasification, and charcoal combustion. The sample is introduced into the preheated reactor, and therefore, a rapid heating, which is well-comparable to the heating in real thermal conversion processes, can be achieved. The test conditions for all test runs performed with the lab-scale reactor are defined in Figure 1. The following measurements/analyses were performed during each of the combustion test runs: (1) Mass decrease of the fuel over time (balance). (2) Concentrations of flue gas species over time: (i) Determination of CO2, H2O, CO, CH4, NH3, HCN, N2O, and basic hydrocarbons was performed with a multi-component Fourier transform infrared (FTIR) spectroscopy device (Ansyco Series 447). (ii) The O2 and H2 concentrations were measured with a multicomponent gas analyzer (Rosemount NGA 2000). This device uses paramagnetism as a measurement principle for O2 detection and a thermal conductivity detector (TCD) for H2 detection. (iii) The amount of total hydrocarbons (CxHy) was determined with a flame ionization detector (FID, Bernath Atomic 3005), which detects organic compounds by ionization in a burning H2 flame. (iv) Wideband λ sensor (O2). (v) A nitric oxide analyzer (ECO Physis CLD 700 El ht), which uses the principle of chemiluminescence, has been used for NO and NO2 detection. (3) Temperature measurements over time: (i) Five thermocouples in the fuel bed (three different heights NiCr−Ni). (ii) Thermocouples in the gas phase (NiCr−Ni). (4) Analysis of the biomass fuel used and residual ash produced (see section 3). The data of the fuel analysis and the residual ash analysis as well as the weight measurements of the fuel sample and the ash sample can be 382

dx.doi.org/10.1021/ef201282y | Energy Fuels 2012, 26, 380−390

Energy & Fuels

Article

used to calculate the elemental release to the gas phase. This is performed by calculating the mass balance for every relevant element as well as the total ash. The lab-scale reactor has been designed to represent the burning conditions of a biomass fuel layer on a grate as well as possible (Figure 1, right). This approach is valid if diffusional transport and mixing effects on the grate can be neglected in comparison to the transport of the fuel along the grate. The validation has been achieved in previous research, which has shown that the fuel transport along the grate can be fluidically characterized by a plug flow in good approximation.10,11 2.5. Definition of Fuel Indexes. On the basis of the evaluation of the fuel analysis and test runs as well as a theoretical evaluation, the following fuel indexes have been investigated and are presented in the following: (1) N concentration in the fuel as an indicator for NOx emissions, (2) sum of K, Na, Zn, and Pb as an indicator regarding aerosol emissions (fine particles smaller than 1 μm) and deposit buildup, (3) molar Si/K ratio for an estimation of the K release from the fuel to the gas phase, (4) molar 2S/Cl ratio for the prediction of the risk of high-temperature corrosion, (5) molar (K + Na)/[x(2S + Cl)] ratio for the prediction of the gaseous emissions of SOx and HCl, and (6) the prediction of the ash-melting temperatures with the molar Si/(Ca + Mg) and molar (Si + K + P)/(Ca + Mg) ratios. The definitions of the above-mentioned indexes are based on theoretical considerations. With the exception of the molar Si/K ratio and the factor x occurring in the index (K + Na)/[x(2S + Cl)], all indexes were validated using results of pilot- and real-scale combustion tests. The factor x is a function of the release of K, Na, S, and Cl from the fuel to the gas phase, which has been derived from lab-scale reactor tests (see Figure 1), because of the well-defined conditions provided by this unit. To evaluate the index of the molar Si/K ratio, data from the lab-scale reactor have also been used.

ultracentrifugal mill equipped with a titanium rotor and screen to a particle size 1, very small HCl and SOx emissions have to be expected, because most S and Cl will be bound in the ash. If the value of the index is clearly 90 wt %. 4.6. Molar 2S/Cl Ratio as an Indicator for HighTemperature Corrosion Risks. With regard to corrosion in biomass-fired boilers, in particular, three mechanisms are relevant:25,26 (1) the direct attack of gaseous HCl or Cl2 to heat-exchanger surfaces, (2) the formation of alkali sulfate and/ or alkali chloride melts, which dissolve the protective oxide layer of the heat-exchanger surface, and (3) the sulfation of alkali metal or heavy metal chlorides in the tube near the deposition layer. From this mechanism, Cl is released, which subsequently attacks the tube surface (so-called active oxidation). Among the three mechanisms mentioned, active oxidation is the most relevant mechanism regarding high-temperature corrosion in boilers.25 As already explained in sections 4.4 and 4.5, S and Cl show almost constant release ratios for different biomass fuels. Both elements are relevant for aerosol and deposit formation because, in the gas phase, they form alkaline sulfates and alkaline chlorides, which subsequently form particles or condense on heat-exchanger surfaces. Therefore, a link between the 2S/Cl ratio in the fuel and the respective aerosol deposits formed prevails. In Figure 7, data regarding the 2S/Cl ratio in

Figure 7. Dependency between the molar ratios of 2S/Cl in fuels and aerosol particles. Explanations: no statistical significance. 386

dx.doi.org/10.1021/ef201282y | Energy Fuels 2012, 26, 380−390

Energy & Fuels

Article

It can be assumed that Na shows a comparable release behavior to the gas phase as K. Therefore, the values of K release can be used for calculating the factor x of the molar (K + Na)/[x(2S + Cl)] ratio. An averaged release for S and Cl (90 wt %) is divided by the average K release to obtain the factor x. The calculated values are summarized in Table 1.

On the basis of the results from pilot, real, and lab reactor combustion tests with different biomass fuels, the release rates for K were calculated. The release was quantified by a mass balance based on weight measurements and chemical analysis of the fuel and ash obtained. The formula of the release rate for pilot- and real-scale combustion tests is shown in eq 1, and the formula of the release rate for lab-scale combustion tests is shown in eq 2.

Table 1. Experimentally Determined K-Release Rates for Different Biomass Fuels and the Resulting Factor x

releaseK,wt % = (1 − (cK,bottom ash(db) (mbottom ash(db) + mfurnace ash(db)

hardwood chips softwood chips waste wood bark A. donax straw maize residues grass pellets decanter and rapeseed press cake residues of starch production

+ mboiler ash(db))) /(cK,fuel(db)m fuel(db))) × 100

(1)

The incoming mass flow of fuel is mfuel(db). The mass flow that is not released to the gas phase is the sum of the remaining ash fraction on the grate mbottom ash(db) (kg/h, db) and the mass of entrained particles from the fuel bed. The mass flows (kg/h, db) that are entrained from the fuel bed and separated in the boiler are mfurnace ash(db) and mboiler ash(db). The chemical composition of the fuel is cK,fuel(db) (mg/kg, db), and the chemical composition for the ash sample is cK,bottom ash(db). It is assumed that the entrained ash particles [mfurnace ash(db) and mboiler ash(db)] have the same composition as the bottom ash. This of course neglects the mass of ash-forming vapors, which end up on the surface of coarse fly ash particles (e.g., by condensation or surface reactions) in the boiler (e.g., K2SO4 and KCl). However, these processes have only a minor influence on the total mass of entrained particles.

(

)

standard deviation (wt %)

factor x

32.8 24.9 18.5 6.8 18.5 20.1 12.1 13.5 8.6

12.0 8.1 3.2 1.3 4.6 2.6 1.7 1.3 1.6

2.7 3.6 4.9 13.2 4.9 4.5 7.4 6.7 10.5

10.8

2.5

8.4

For biomass fuels for which the release behavior has not been investigated thus far, the release can be determined with a labscale reactor experiment, such as, for instance, described in section 2.4. On the basis of the results of such an experiment, a meaningful assumption for the K release can be made and, thus, the factor x can be calculated. The estimated values from Table 1 can be used for the calculation of the index (K + Na)/[x(2S + Cl)]. Figure 9 shows the correlation of the molar (K + Na)/[x(2S + Cl)] ratio versus the SOx emissions and the conversion of fuel S to S in SOx. From Figure 9, it can be seen that, with a decreasing molar (K + Na)/[x(2S + Cl)] ratio, SOx emissions are increasing. The SOx emissions of the different fuels can be categorized into (1) a low SOx emission range (500 mg of SOx N−1 m−3, with dry flue gas, at 13 vol % O2), e.g., grass pellets, a mixture of decanter and rapeseed press cake, and residues of starch production with high S content. In the second diagram of Figure 9, the molar (K + Na)/ [x(2S + Cl)] ratio is plotted against the conversion of fuel S to S in SOx. The same trend of the molar (K + Na)/[x(2S + Cl)] ratio and the conversion of fuel S to SOx can be observed. SOx emissions are negligible for a fuel molar (K + Na)/[x(2S + Cl)] ratio bigger or close to 0.5. This index is suitable to estimate the SOx emission range to be expected. Figure 10 shows the correlation of the molar (K + Na)/ [x(2S + Cl)] ratio versus HCl emissions and the conversion of fuel Cl to Cl in HCl. No clear correlation between the molar (K + Na)/[x(2S + Cl)] ratio and the HCl emissions can be seen. For a molar ratio of (K + Na)/[x(2S + Cl)] < 0.5, the HCl emissions are varied from 0 to 110 mg N−1 m−3, with dry flue gas, at 13 vol % O2. The reason for this strong scattering in comparison to S is not yet understood. Further investigations are needed. In our case, it can be seen from Figure 10 that, with a molar ratio of (K +

releaseK,wt % = 1 − (cK,output(db)moutput(db)) /(cK,input(db)minput(db)) × 100

average K release (wt %)

(2)

In equation (2) minput(db) is the mass (g, db) of the sample fed to the reactor, cK,input(db) is the concentration of K in the fuel (mg/kg, db), moutput(db) is the mass (g, db) of the ash residual after the experiment in the reactor and cK,output(db) is the concentration of K in the ash residual (mg/kg, db). The mass of entrained fly ash particle is negligible because of the much lower air velocities in the fuel bed compared to pilot- or realscale applications. The determined release rates are shown in Figure 8.

Figure 8. K release for different biomass fuels during combustion. 387

dx.doi.org/10.1021/ef201282y | Energy Fuels 2012, 26, 380−390

Energy & Fuels

Article

Figure 9. Molar (K + Na)/[x(2S + Cl)] ratio versus SOx emissions and the conversion of fuel S to SOx. Explanation: both correlations are statistically significant (p < 0.05).

is a linear correlation given between the molar (Si + P + K)/ (Ca + Mg) ratio and the ash-sintering temperature.

Na)/[x(2S + Cl)] bigger or close to 0.5, the HCl emissions are very low. Therefore, this index can also be used to make a first estimation of the gaseous HCl emissions to be expected. When using this index, it has to be noted that the factor x is based on first estimations regarding the release of K and Na. Further investigations are recommended to ensure the alkali metal release described. 4.9. Indicators for Ash-Melting Problems. It is generally well known that Ca and Mg increases the ash-melting temperature, while Si in combination with K decreases the ash-melting temperature.9,27,28 The molar Si/(Ca + Mg) ratio9 can therefore provide first information about ash-melting tendencies in ash systems dominated by Si, Ca, Mg, and K. However, for P-rich systems (e.g., grass pellets), this correlation is not valid (see Figure 11). As seen in Figure 11, the ash-sintering temperature drops below 1100 °C as soon as the Si/(Ca + Mg) ratio exceeds 1. It can also be seen that there is a linear correlation of the molar Si/(Ca + Mg) ratio with the ash-sintering temperature, except for grass pellets that show a low ash-sintering temperature because of the high P concentration. From the ternary-phase diagrams for CaO−K2O−P2O5 and MgO−K2O−P2O5,23 it can be derived that, at constant K2O/ P2O5 ratios, the melting point increases with increasing CaO and MgO concentrations. It can be concluded that CaO and MgO increase the ash-sintering temperature, whereas K2O and P2O5 decrease the sintering temperature. In combination with Si, a modified index (Si + P + K)/(Ca + Mg) can be introduced (see Figure 12). With this index, also for P-rich fuels, a prediction regarding the ash-melting behavior is possible. There

5. SUMMARY AND CONCLUSION The application of fuel indexes as a characterization tool provides a good basis for a quick pre-evaluation of combustionrelated problems that may arise. This new and advanced fuel characterization method can therefore be applied to support decision making concerning the application of new biomass fuels or fuel blends in existing combustion plants, as well as for the preliminary design and engineering of new combustion plants, which should be tailored to the needs of a specific fuel or fuel mixture. During this first decision phase regarding the applicability of a certain fuel, time-consuming and expensive combustion tests can therefore be saved. The investigation of the fuel indexes has been performed for a broad spectrum of biomass fuels ranging from different types of wood to herbaceous and agricultural biomass, as well as industrial biomass residues. The data for the investigation are based on test runs from fixed-bed lab-, pilot-, and real-scale combustion systems, and therefore, the indexes derived in this work are applicable for grate combustion plants. Fuel indexes, which allow for accurate qualitative predictions, are the N content, the molar 2S/Cl ratio, and the molar ratio of (Si + P + K)/(Ca + Mg). Indexes that can be applied with some restrictions regarding other constraints are the sum of K, Na, Zn, and Pb and the molar ratios of (K + Na)/[x(2S + Cl)]) and Si/(Ca + Mg). The N content generally determines the potential for NOx emission formation. A correlation between the conversion rates 388

dx.doi.org/10.1021/ef201282y | Energy Fuels 2012, 26, 380−390

Energy & Fuels

Article

Figure 10. Molar (K + Na)/[x(2S + Cl)] ratio versus HCl emissions and the conversion of fuel Cl to HCl. Explanation: both correlations are not statistically significant.

Figure 11. Molar Si/(Ca + Mg) ratio versus ash-sintering temperature for different biomass fuels. Explanation: if grass pellets are excluded, the correlation is statistically significant (p < 0.05); ash-sintering temperature according to prCEN/TS 15370-1.

Figure 12. Molar (Si + P + K)/(Ca + Mg) ratio versus ash-sintering temperature for different biomass fuels. Explanation: the correlation is statistically significant (p < 0.05).

potential for aerosol and deposit formation, which increases with rising values of this index. Moreover, the molar ratio of (K + Na)/[x(2S + Cl)], where x represents the ratio of the average K and Na release in relation to the average S and Cl release from the fuel to the gas phase, provides a first prediction concerning the expected HCl and SOx emissions. For values >0.5, it is very likely that most S and Cl will be embedded in the ashes. The molar ratio of Si/(Ca + Mg) can provide first information about ash-melting tendencies in systems dominated by Si, Ca, Mg, and K (decreasing melting temperatures with an increasing value). The molar ratio of (Si + P + K)/(Ca + Mg) is also valid for P-rich fuels and is therefore an index that can generally be applied regarding the pre-evaluation of ashmelting tendencies.

of fuel N to NOx emissions has been derived, showing a nonlinear correlation of the NOx emissions with rising fuel N content. When this correlation is applied, a first estimation of the NOx emissions can be made. During combustion, easily volatile and semi-volatile elements (K, Na, S, Cl, Zn, and Pb) are partly released from the fuel to the gas phase, where they undergo chemical reactions and finally contribute to problems concerning emissions, deposit formation, and corrosion. With a decreasing molar 2S/Cl ratio, for instance, the amount of alkaline chlorides in ash deposits on heat-exchanger surfaces increases. According to the literature,25 a molar ratio of 2S/Cl < 8 increases the risk of high-temperature Cl corrosion. The sum of K, Na, Zn, and Pb (in mg/kg of dry fuel) describes the 389

dx.doi.org/10.1021/ef201282y | Energy Fuels 2012, 26, 380−390

Energy & Fuels

Article

(14) Glarborg, P.; Jensen, A. D.; Johnsson, J. E. Fuel nitrogen conversion in solid fuel fired systems. Prog. Energy Combust. Sci. 2003, 29, 89−113. (15) Weissinger, A. Experimentelle Untersuchungen und reaktionskinetische Simulationen zur NOx-Reduktion durch Primärmaßnahmen bei Biomasse-Rostfeuerungen. Ph.D. Thesis, Graz University of Technology, Graz, Austria, 2002 (16) Brunner, T. Aerosols and coarse fly ashes in fixed-bed biomass combustion. Ph.D. Thesis, Graz University of Technology, Graz, Austria, 2006. (17) Christensen, K. The formation of submicron particles from the combustion of straw. Ph.D. Thesis, Department of Chemical Engineering, Technical University of Denmark, Lyngby, Denmark, 1995; ISBN: 87-90142-04-7. (18) Jöller, M. Modelling of aerosol formation and behaviour in fixed-bed biomass combustion systemsInfluencing aerosol formation and behaviour in furnaces and boilers. Ph.D. Thesis, Graz University of Technology, Graz, Austria, 2008. (19) Knudsen, J.; Jensen, P.; Dam-Johansen, K. Transformation and release to the gas phase of Cl, K, and S during combustion of annual biomass. Energy Fuels 2004, 18, 1385−1399. (20) Novaković, A.; Van Lith, S.; Frandsen, F.; Jensen, P.; Holgersen, L. Release of potassium from the systems K−Ca−Si and K−Ca−P. Energy Fuels 2009, 23, 3423−3428. (21) Müller, M.; Wolf, K. J.; Smeda, A.; Hilpert, K. Release of K, Cl and S species during co-combustion of coal and straw. Energy Fuels 2006, 20, 1444−1449. (22) Boström, D.; Broström, M.; Skoglund, N.; Boman, C.; Backman, R.; Ö hman, M.; Grimm, A. Ash transformation chemistry during energy conversion of biomass. Proceedings of the Impact of Fuel Quality on Power Production and Environment; Puchberg, Austria, Sept 23−27, 2010; pp 1−20. (23) Sandström, M. Structural and solid state EMF studies of phases in the CaO−K 2O−P 2O5 system with relevance for biomass combustion. Ph.D. Thesis, Luleå University of Technology, Luleå, Sweden, 2006; ETPC Report 06-06, ISSN: 1653-0551, ISBN: 917264-149-5. (24) Risnesa, H.; Fjellerupb, J.; Henriksenc, U.; Moilanend, A.; Norbye, P.; Papadakisb, K.; Posseltf, D.; Sorensenb, L. H. Calcium addition in straw gasification. Fuel 2003, 82, 641−651. (25) Salmenoja, K. Field and laboratory studies on chlorine-induced superheater corrosion in boilers fired with biofuels. Ph.D. Thesis, Faculty of Chemical Engineering, Abo Akademi, Abo/Turku, Finland, 2000. (26) Obernberger, I.; Brunner, T. Depositionen und Korrosion in Biomassefeuerungen. Tagungsband zum VDI-Seminar 430504 “Beläge und Korrosion in Großfeuerungsanlagen; Göttingen, Germany, 2004. (27) Ö hman, M.; Boman, C.; Hedman, H.; Nordin, A.; Boström, D. Slagging tendencies of wood pellet ash during combustion in residential pellet burners. Biomass Bioenergy 2003, 27, 585−596. (28) Baxter, L.; Miles, T.; Miles, T. Jr.; Jenkins, B.; Milne, T.; Dayton, D.; Bryers, R.; Oden, L. The behavior of inorganic material in biomassfired power boilers: Field and laboratory experiences. Fuel Process. Technol. 1998, 54, 47−78.

Presently, ongoing work is focusing on the further improvement of existing fuel indexes. For the prediction of HCl and SOx emissions, a more detailed investigation of the K release to the gas phase and also possible Ca interactions is necessary. To investigate the K release from biomass fuels in more detail, interactions of Si and P with K as well as the combustion temperature need to be considered. However, more research on this special issue is needed.



AUTHOR INFORMATION

Corresponding Author

*Telephone: +43-(0)-316-481300-73. Fax: +43-(0)-316481300-4. E-mail: [email protected].



ACKNOWLEDGMENTS This paper is the result of a project carried out in cooperation with BIOENERGY 2020+ (a research center within the framework of the Austrian COMET program), which is funded by the Republic of Austria as well as the federal provinces of Styria, Lower Austria, and Burgenland.



REFERENCES

(1) Gustavson, L.; Holmberg, J.; Dornburg, V.; Sathre, R.; Eggers, T.; Mahapatra, K.; Marland, G. Using biomass for climate change mitigation and oil use reduction. Energy Policy 2007, 35, 5671−5691. (2) McKendry, P. Energy production from biomass (Part 1): Overview of biomass. Bioresour. Technol. 2002, 83, 37−46. (3) Hoogwijk, M.; Faaij, A.; van den Broek, R.; Berndes, G.; Gielen, D.; Turkenburg, W. Exploration of the ranges of the global potential of biomass for energy. Biomass Bioenergy 2003, 25, 119−133. (4) Lewandowski, I.; Scurlock, J. M. O.; Lindvall, E; Christou, M. The development and current status of perennial rhizomatous grasses as energy crops in the U.S. and Europe. Biomass Bioenergy 2002, 25, 335−361. (5) Chouchene, A.; Jeguirim, M.; Khiari, B; Zagrouba, F.; Trouvé, G. Thermal degradation of olive solid waste: Influence of particle size and oxygen concentration. Resour., Conserv. Recycl. 2010, 54, 271−277. (6) Bryers, R. W. Fireside slagging, fouling, and high-temperature corrosion of heat-transfer surfaces due to impurities in steam-raising fuels. Prog. Energy Combust. Sci. 1996, 22, 29−120. (7) Visser, H. J. M. The Influence of Fuel Composition on Agglomeration Behaviour in Fluidised-Bed Combustion; Energy Research Centre of the Netherlands (ECN): Petten, The Netherlands, 2004; ECN-C-04-054. (8) Gilbe, C.; Ö hman, M.; Lindström, E.; Boström, D.; Backman, R.; Samuelsson, R.; Burval, J. Slagging characteristics during residential combustion of biomass pellets. Energy Fuels 2008, 22, 3536−3543. (9) Lindström, E.; Ö hman, M.; Backman, R.; Boström, D. Influence of sand contamination on slag formation during combustion of wood derived fuels. Energy Fuels 2008, 22, 2216−2220. (10) Weissinger, A.; Fleckl, T.; Obernberger, I. Application of FT-IR in-situ absorption spectroscopy for the investigation of the release of gaseous compounds from biomass fuels in a laboratory scale reactor Part II: Evaluation of the N-release as functions of combustion specific parameters. Combust. Flame 2004, 137, 403−417. (11) Stubenberger, G.; Scharler, R.; Obernberger, I. Nitrogen release behaviour of different biomass fuels under lab-scale and pilot-scale conditions. Proceedings of the 15th European Biomass Conference and Exhibition; Berlin, Germany, May 7−11, 2007; ISBN: 978-88-8940759-X, ISBN: 3-936338-21-3, pp 1412−14120. (12) Bärenthaler, G.; Zischka, M.; Hanraldsson, C.; Obernberger, I. Determination of major and minor ash-forming elements in solid biofuels. Biomass Bioenergy 2006, 983−997. (13) Keller, R. Primärmaßnahmen zur NOx-Minderung bei der Holzverbrennung mit dem Schwerpunkt der Luftstufung. Forschungsbericht 18: Laboratory for Energy Systems; ETH Zürich: Zürich, Switzerland, 1994.



NOTE ADDED AFTER ASAP PUBLICATION An addition sign was added between Ca and Mg in the fourth to last paragraph of the Background and Scope section of the version of this paper published December 16, 2011. The correct version published December 20, 2011.

390

dx.doi.org/10.1021/ef201282y | Energy Fuels 2012, 26, 380−390